Detecting Selection Signatures from Population Differentiation: Results from the SheepHapMap and 1000 Bull Genomes Datasets

The genetics of present day livestock populations has been shaped by evolution in particular through genetic drift and selection, from domestication to modern genetic improvement. Identifying genes that have been underlying adaptation of these populations is one of the possible approaches to deciphering the genetic basis of livestock phenotypes that have been under selective pressure. A common strategy to identify these selection signatures is based on identifying regions where populations exhibit particularly high genetic differences, i.e.  more than expected from drift alone. We recently improved on the classical statistical approaches for detecting highly differentiated genome regions by developping an haplotype-based approach that also accounts for the phylogeny of populations. 

We will present applications of this method on two different datasets: the SheepHapMap dataset, a world-wide survey of sheep diversity through the genotyping of 50K SNPs in many populations and preliminary data of the 1000 Bull Genomes dataset, based on complete resequencing of individuals from four cattle breeds. These applications illustrate the advantages and limitations of differentiation-based approaches for detecting selection signatures and provide insight on the potential achievements of such approaches when applied to complete sequence data.